This study investigates the spatial variability of some remote sensing indices representing built-up areas, vegetation, bareness, and urban heat island (UHI), based on time-series Landsat TM/ETM+ and OLI/TIRS datasets archived for 1980–2020 period from the US Geological Survey’s website for Kumasi Metropolitan Area in Ghana. Modules for Land Use Change Evaluation (MOLUSCE) and Cellular Automata Artificial Neural Network (CA-ANN) algorithms and simulations in QGIS were used to predict future changes (2020–2050) for land-use systems in Kumasi. Findings revealed urbanization/built-up areas (+108.02%) contributed massively to the decline of forest areas (?93.34%) and farmlands/shrubs (?31.53%), thereby making Kumasi lose its once critical green position as the “Garden City of West Africa.” UHI moderately and strongly correlated positively against built-up (R2=0.78, p < 0.0001) and bareness (R2=0.96, p < 0.0001) indices, respectively. By contrast, UHI showed a statistically significant inverse relationship with the vegetative index (R2=0.97, p < 0.0001). Future land-use scenarios revealed more forests, waterbodies, and farmlands/shrubs will be lost, influencing urban temperature and water supply. The multipurpose advantages of urban green space are ingrained in the grand urban model. Contextually, the Metropolis’s resilience has been hampered by inconsistency in the performance of institutional roles, competition for land ownership rights over green areas, and little investments or prioritization of green spaces. An integrated collaborative governance framework is proposed to unify actions, address power crisis and factors that influence governance of green infrastructure, UHI and land cover change.